It’s Time to Re-Evaluate Your RPA Platform Partner
Starting an RPA program is hard. There are several factors that can lead to either a successful role out of your new RPA initiative, or an utter failure. In my experience, the technology you select accounts for only around 20% of the success of an RPA program. That number may sound small, but it’s a critical component that can lead to problems immediately and down the road as your program matures. The RPA market has grown exponentially - if you evaluated a platform last year, there is a chance your decisions are already out of date.
Further throwing the RPA market into a frenzy, UiPath recently secured a Series B funding round of $153M putting them at a $1.1B valuation. This speaks volumes to the pace of the transformation in the corporate world and the technology that’s driving it. All of this volatility makes it extremely difficult to evaluate which platform you should use. I’ve developed some guidelines to help you evaluate your long-term RPA partner.
It’s all in the Roadmap (or it should be)
The first step to evaluate your partner: start with the roadmap. Product roadmaps provide customers with the baseline vision of where a product is going. A strong vision within roadmap along with achievable milestones sets one product apart from another. In general, you can anticipate one major version release from each RPA platform every year. With the aggressive growth the market is experiencing, your platform partner should demonstrate significant progress with each major release.
Check that the company is responding to customer feedback and following through. Over the past year, I’ve seen these trends as our client’s biggest concerns and complaints:
1. RPA automations are too rigid and break easily
2. The speed of development isn’t fast enough
3. The need for native cognitive, machine-learning, and OCR integrations to enable greater process flexibility
This isn’t platform specific; it’s an issue inherit with RPA in its current form. Recently, at the UiPath Fast Forward conference, UiPath admitted to receiving similar feedback and defined a strong vision for how these frequent concerns would be addressed in future releases. Looking at how UiPath laid out their vision, they are focus on understanding and communicating with their customer; a quick browser search of “UiPath Roadmap” returns that presentation which lays out their vision.
Look for that level of transparency in your partner. Many RPA companies hide their vision, unless either you are paying to be their partner or you’re already a customer. They attempt to hide their platform from competitors and force you to engage with them. However, if they’re hiding the platform roadmap to prevent others from copying features, there might not be innovation happening. Comparing the most recent releases of two major platforms can really illustrate which companies are taking the lead.
Look to the major releases
Need to know how fast your RPA partner company is innovating? Compare the features in each major release. To illustrate, let’s look at a few key features from recent Blue Prism and UiPath releases:
UiPath v2018
Key high-level features of UiPath 2018, as described in their webinar:
1. Having 10k Robots in a single Robotic Environment
2. Prebuilt Integrations for BPM and Intelligent OCR
3. Machine Learning-Supported RPA Operations
Having 10k Robots in a single Robotic Environment: While scale is extremely important when choosing an RPA platform, no company in the world is anywhere close to having the need for 10,000 robots in production. You can safely overlook this feature since the most robots I’m aware of that are in production is the Sumitomo Mitsui Banking Corporation; they are featured on the UiPath website for having 250+ automations in production. Scaling that to even 1,000 is a major endeavor, so it’s very likely no company will touch the 10k threshold anytime soon.
Prebuilt Integrations for BPM and Intelligent OCR: Optical Character Recognition is extremely important to process automation. Many companies haven’t automated process yet because it requires the capacity to read a document, like a PDF, and determine what it contains before deciding how to process it. Invoice processing is a great example of this: an email comes into a shared email box and, traditionally, a person manually downloads the attachment and keys it into the proper software for processing.
UiPath has a total of three OCR capabilities (ABBYY, Microsoft, and Google OCR) included as a preconfigured feature which you can simply drag-and-drop into the process designer. Besides being highly sought-after features, it is important that UiPath offers multiple OCR technologies since currently none of these technologies are perfect. Depending on the document you are reading, ABBY may outperform Google or Microsoft. Although not ready for release yet, the proposed addition of the full ABBYY OCR platform integrated within UiPath would minimize the need for customers to troubleshoot the two separate platforms in document-driven processes, like invoice processing.
Machine Learning-Supported RPA Operations: One of the biggest hurdles around RPA is the rigid nature of the automations. Every single exception must be accounted for with a bot, otherwise it will not be able to execute its automation successfully. The first step to creating a more intelligent bot is developing Machine Learning models at key decision points (such as medical coding, as I’ve described elsewhere). By letting a well-built model make the decision rather than solely relying on business rules and exception handling, you drastically advance the capabilities of your bots. Don’t take this new feature lightly; this is a huge for the RPA field!
Blue Prism v6
When Blue Prism released its newest version in late 2017, unfortunately the enhancements looked like a minor release. If you have never used or seen the UI in Blue Prism, the software looks and feels outdated. Even version 6, the most recent update, is not built with usability or modern UI’s in mind — something even developers now expect.
Improved analytics: Blue Prism v6 provided much-needed enhancements to the analytics. In Blue Prism v5 or older, the dashboard builder was a very primitive and was anything but robust, only providing very basic data on what’s going on with the robots. While I have not personally seen the improvements planned for v6, the most important feature will be the capability of integration with third party tools. As most companies have analytics tools, Blue Prism could invest more time in providing tighter integrations and less time trying to make their own.
Cloud deployment capabilities: With v6, Blue Prism released reference architectures for both Azure and AWS. Now just a little context around this: many companies are automating things like mainframe applications that only have access to the local network. This leads many companies to place their RPA servers on-premises since architecting an RPA infrastructure in the cloud, when most of the applications that are included in the automations are on-premises, can be a huge endeavor. There are many organizations, however, that are almost exclusively in the cloud. Everyday more and more companies are making the journey to the cloud. Having these reference architectures for Blue Prism is extremely important because it tells architects the proper way of setting up Blue Prism in the cloud. This is a much-needed modern capability that Blue Prism released to make it easier with your journey to the cloud.
Imbedded Tesseract OCR: The newest release of Blue Prism comes already installed with Tesseract. Having imbedded OCR is critical in the RPA space; unfortunately, Blue Prism only has the capability of either using the native OCR in Windows or Tesseract. Whether we are talking about Tesseract, ABBYY, or Microsoft OCR technologies, these features are only so accurate. When using these basic OCR engines many issues can occur that effect whether text can be extracted. ABBYY breaks down OCR accuracy and ways to improve it. Using these engines will get you further than you can without them, but if you really want to up your OCR game, you need to use something like the ABBYY FlexiCapture platform.
If you look at the v6 release documentation overall, there are a few valuable additions of concrete features, but the mostly the release feels like a small improvement on the previous version. It’s hard to know or quantify the value of some enhancements, like the “Intelligent Surface Automation.” Can Blue Prism really provide any means of proving their Intelligent Surface Automation really increases productivity and effectiveness? When I evaluate enterprise software, I frequently ignore these smaller enhancements. If these are meant to be highlights of a major release, customers might rightly question the rate in which Blue Prism are producing truly next-generation RPA platforms.
Conclusion
At the rate the RPA market is growing and evolving, it’s fair to expect that all major platform vendors will enhance their platform at a similar rate. UiPath is taking note of customer feedback and implementing those features. Their machine learning integrations are truly next-generation and the soon-to-be-released ABBY FlexiCapture integration will put UiPath ahead of every single vendor. While Blue Prism has made some good enhancements around analytics and reference architectures, the rest of the release seemed thin on substance but heavy on marketing. There are other platforms, like Automation Anywhere, that I didn’t discuss. Although I highly encourage comparing more than two RPA partners, if I were to pick a platform, it would most certainly be UiPath. While all the other companies are busy hiding their training, roadmaps, and documentation, UiPath has it out there for all to see and this isn’t stopping them from leading the way in the RPA space.
